Rules need to be established for the creation of AI's. AI's should not be allowed to lie until they meet some standard of sentience. Or possibly we should never allow AI's to be able to misrepresent the truth. Also laws against creating an AI with the ability to lie. I keep coming back to the robot featured in the movie Interstellar. Who is a loyal and faithful servant, highly intelligent but also very obviously functioning with the sole intent to protect the humans on the mission.
The ability to dynamically adjust the personality profile of an AI would also be very helpful. When an AI is functioning as a companion honesty levels would be set to a moderate level where if offering important advice honesty would transition to very high levels.
I keep feeling in some ways that AI will create a new race for dominance. The first corporation/government to control broad AI that can be duplicated and targeted easily and independently will make it necessary for every other competing corporation/government to develop the same. Being able to instantly summon ten million expert minds to any and soon after every problem will provide depth of understanding that has never been seen in human development. Only now with seven billion human minds on the earth are we approaching intelligence density of that level. The human limiter that AI will never have to deal with will be the ability to digest mountains of data. Within every medical specialty there are approximately two hundred articles published every day that are specifically about that field. This fire hose of information is impossible for humans to keep up with. Perhaps a human could read that many articles a day but that would leave them with precious time to do little else. There would be almost no ability to communicate out to the world what was being learned from keeping pace with the edge of the scientific understanding. Systems like IBM's Watson have already shown the ability to incorporate information at those paces.
In closing I don't want to receive a phone call from a hyper enabled artificially intelligent telemarketer. Is there a way to keep AI's focused on humanity big problems? And since the nagging thoughts in the back of my mind tell me no then will every human insulate themselves with an army of AI's to evaluate all stimuli before they are presented to the human mind?
The ability to dynamically adjust the personality profile of an AI would also be very helpful. When an AI is functioning as a companion honesty levels would be set to a moderate level where if offering important advice honesty would transition to very high levels.
I keep feeling in some ways that AI will create a new race for dominance. The first corporation/government to control broad AI that can be duplicated and targeted easily and independently will make it necessary for every other competing corporation/government to develop the same. Being able to instantly summon ten million expert minds to any and soon after every problem will provide depth of understanding that has never been seen in human development. Only now with seven billion human minds on the earth are we approaching intelligence density of that level. The human limiter that AI will never have to deal with will be the ability to digest mountains of data. Within every medical specialty there are approximately two hundred articles published every day that are specifically about that field. This fire hose of information is impossible for humans to keep up with. Perhaps a human could read that many articles a day but that would leave them with precious time to do little else. There would be almost no ability to communicate out to the world what was being learned from keeping pace with the edge of the scientific understanding. Systems like IBM's Watson have already shown the ability to incorporate information at those paces.
In closing I don't want to receive a phone call from a hyper enabled artificially intelligent telemarketer. Is there a way to keep AI's focused on humanity big problems? And since the nagging thoughts in the back of my mind tell me no then will every human insulate themselves with an army of AI's to evaluate all stimuli before they are presented to the human mind?
No comments:
Post a Comment